Independent employee
Owns a dedicated service account and role: granted system permissions like a real employee, executing tasks under its own identity with full execution records.
Capabilities
The NexusClaw capability map
Every core capability of the AI-native CRM and agent governance layer: identity and authority, the growth loop, governance and audit, deployment topologies, CLI and developer workflows, IM integrations, and mobile.
Digital Workforce
AI digital employees: two identities, clear authority
Every digital employee has an explicit authorization subject — identity determines where permissions come from.
Assistant
Acts within the initiating user’s permissions for each task — authority is inherited, scoped, and traceable.
Execution engine
ReAct-style stepwise execution with tool calls, human approval checkpoints, handoffs, and result write-backs; unauthenticated calls are denied by default.
Growth Loop
The growth loop: human corrections become rollback-able assets
Digital employees get more valuable over time — because every correction is captured. Improvements pass a release gate; humans stay in the loop.
Behavior learning pipeline
Human corrections, tool failures, knowledge gaps, and SOP drift are recognized and structured into reviewable improvement recommendations (prompts, SOPs, guardrails, model routing, eval suites).
Release gate
Nothing reaches production behavior without tests, simulation, human approval, and regression against baseline — there is no path that bypasses the gate.
Versioned and rollback-able
Prompt versions, SOPs, guardrail rules, model routing, and eval suites are exportable, versioned, rollback-able digital assets — not disposable config files.
CRM behavior distillation
Human CRM operations are distilled into SOPs and injected into knowledge retrieval — your team’s best practice becomes how digital employees work.
Governance
Governance and audit: the enterprise control plane
From permissions to traces, retrieval to budgets — every step is auditable.
Execution audit chain
Agent executions, reasoning steps, tool calls, approval decisions, and outbox events are fully recorded — replayable and attributable.
Permission-aware retrieval
Knowledge retrieval (RAG) checks workspace, role, and object-read permissions per candidate — deny by default; out-of-scope content never enters context.
Guardrails and budgets
Guardrails bound behavior; model usage is metered per workspace with visible cost and configurable ceilings.
Deployment
Deployment: from managed SaaS to data-stays-on-prem
One kernel, three topologies — choose by data boundary and compute.
Managed SaaS
Ready on activation; database, cache, and object storage are managed by the platform.
Hybrid (cloud + AI PC local models)
The system runs in the cloud while model inference runs on your own AI PC / intranet machine: models and sensitive data stay inside your network.
Private / open-source self-host
Single-node private delivery, or self-host the open-source Community edition (AGPL-3.0-only) with Docker Compose — full stack measured ready in 35s (2026-08-15, single-host Docker).
Developers
CLI and developer workflows
The full digital-employee lifecycle, from your terminal.
nexus CLI
npm install --global @nexusclawhq/cli: org login, workspace and metadata management, deployment and release flows — credentials encrypted locally.
workforce lifecycle commands
nexus workforce * covers learning loops, cognition, config writes, evals, ops, orchestration, and observation — run your digital workforce without the UI.
Cross-machine DX
The same CLI works across your laptop, CI, and servers with consistent auth and fail-closed target resolution.
Integrations
IM integrations and mobile
Digital employees live where your team already works.
WeCom / Feishu / DingTalk
Headless integrations with full signature, encryption, and gateway contract alignment — digital employees take tasks and return results inside your existing IM.
Mobile H5
A mobile work surface for approvals, review, and light operations.
Open object model
A metadata-driven object/field/permission system — define custom business objects under the same governance chain.
Scope note: the above reflects currently delivered capabilities. The growth loop is a human-gated governance loop (human corrections and real outcomes consolidated through a release gate) and is not an autonomous self-evolution promise. The open-source Community edition (AGPL-3.0-only) ships the governance-kernel subset — see the GitHub repository.